1. Field
The application generally relates to controlling a fabrication cluster by using profile parameters determined using an optical metrology system and, more particularly, to a method and a system for enhancing the accuracy of determining profile parameters of a sample structure on a workpiece using tracing-based libraries and machine learning systems.
2. Related Art
Optical metrology involves directing an incident beam at a sample structure on a workpiece, measuring the resulting diffraction signal, and analyzing the measured diffraction signal to determine various characteristics of the sample structure. The workpiece can be a wafer, a substrate, photomask or a magnetic medium. In manufacturing of the workpieces, periodic gratings are typically used for quality assurance. For example, one typical use of periodic gratings includes fabricating a periodic grating in proximity to the operating structure of a semiconductor chip. The periodic grating is then illuminated with an electromagnetic radiation. The electromagnetic radiation that deflects off of the periodic grating are collected as a diffraction signal. The diffraction signal is then analyzed to determine whether the periodic grating, and by extension whether the operating structure of the semiconductor chip, has been fabricated according to specifications.
In one conventional system, the diffraction signal collected from illuminating the periodic grating (the measured diffraction signal) is compared to a library of diffraction signals. Each diffraction signal in the library is associated with a hypothetical profile. When a match is made between the measured diffraction signal and one of the diffraction signals in the library, the hypothetical profile associated with the diffraction signal is presumed to represent the actual profile of the periodic grating. The hypothetical profiles, which are used to generate the diffraction signals, are generated based on a profile model that characterizes the structure to be examined. Thus, in order to accurately determine the profile of the sample structure using optical metrology, a profile model that accurately characterizes the sample structure should be used.
With increased requirements for throughput, smaller size of the test structures, smaller spot sizes, and lower cost of ownership, there is greater need to optimize design of optical metrology systems to meet the objectives of the overall application. Current optical metrology systems typically focus on optimizing the variables used in generating the simulated diffraction signals. Accuracy requirements increase as the dimensions of the structures get smaller, for example, as the lithography node goes to 30 nm and smaller. In terms of measurement uncertainty, as the size of the structures get smaller, complicated interactions between the optical metrology tool properties vary in complex ways to affect the accuracy of the measurement. For example, as the lithography node gets smaller, errors associated with critical dimension and sample structure profile extraction are the larger errors to be considered. With a smaller lithography node, the total measurement uncertainty and other characterization of uncertainty need to be considered with all elements that can contribute to the error in the measured signal off the structure. As the size of the structures get smaller, factors that did not substantially affect the measurement accuracy are now making an impact.
Furthermore, assumptions used in modeling the optical metrology tool are no longer sufficient. In order to achieve enhanced accuracy of profile parameters of the structure, considerations regarding the physical optics, geometric optics, beam propagation parameters, and detail analysis of the effect of imperfections of optical elements on the illumination and diffraction beam paths need to be incorporated in the modeling and simulations of the diffraction signal to be used in a profile parameter extraction system.
In ray tracing, when the number of rays is increased, the length of time to create a library and the size of the library can be problematic. As the numerical aperture (NA) is increased, the size of the library is also increased. For training a machine learning systems (MLS), time for training the MLS increases as the number of the training samples used to train the MLS increases. The use of reflection coefficients and Jones matrix further increases the size of the library, thus the time needed for MLS training can be a significant consideration. Some of the simulated diffraction signals can be pre-calculated and averaged before recording or training the MLS. However, the library or MLS is dependent on the specific system calibration parameters. Thus, there is a need for a library with a reasonable size while maintaining the accuracy of the optical metrology system for determining profile parameters of the sample structure. Furthermore, there is a need for an optical metrology system that has the flexibility to handle metrology applications with different system calibration parameters, different ray tracing techniques, different beam propagation parameters, different diffraction metrology signal parameters (metrology signal parameters) while maintaining reasonable response times for integrated or standalone metrology applications.